Metrics that are unserializable will be logged at debug level, but the
rest of the batch will be sent. Unserializable metrics can occur during
normal operation such as if you remove all fields from a metric or the
metric cannot fit within the line size limit.
Using 'g' with -1 precision switches over to scientific notation for too
small of numbers, and setting a larger precision results in larger than
desired representations.
main reasons behind this:
- make adding/removing tags cheap
- make adding/removing fields cheap
- make parsing cheaper
- make parse -> decorate -> write out bytes metric flow much faster
Refactor serializer to use byte buffer
